Lung cancer incidence in cigarette smokers: further analysis of Doll and Hill's data for British physicians

Abstract

Doll's analysis of lung cancer incidence (mortality) in cigarette smokers is refined in more detail in this paper. His conclusion that incidence is approximately proportional to rate of smoking and the fifth power of years of smoking is shown to hold in each of several age and dose groups giving additional evidential support. The data are equally well fitted by a Weibull distribution and by a lognormal distribution with a constant geometric standard deviation. There is, however, a trend away from these fits which needs to be examined in other studies.